Project Summary G-protein-coupled receptors (GPCRs) are the largest superfamily of human membrane proteins and serve as primary targets of about 1/3 of currently marketed drugs. Four subtypes of adenosine receptors, the A1, A2A, A2B, and A3, mediate a broad range of physiological functions. They have emerged as important therapeutic targets for treating cardiac ischemia, neuropathic pain and cancer. During function, the A1 and A3 receptors bind the Gi/o proteins, while the A2A and A2B receptors bind the Gs proteins. Moreover, the GPCR?G protein interactions are modulated by allosteric ligands. These ligands bind to a putative extracellular site of adenosine receptors, which exhibit divergent sequences and conformations. In contrast to traditional agonists that target at the highly conserved adenosine-binding site and often cause off-target side effects, allosteric modulators have emerged as promising candidates as selective GPCR drugs. To date, adenosine receptors are the sole subfamily of GPCRs that have X-ray or cryo-EM structures determined in complex with distinct G proteins. Although these structures provide valuable insights into the GPCR?G protein interactions, they are rather static images of the protein complexes. Current limitations include: (1) It remains unknown how the flexible GPCRs and G proteins dynamically recognize each other. (2) The determinants of specific GPCR?G protein interactions remain unclear. (3) The structural basis and mechanism of allosteric modulator binding in the adenosine receptors remain elusive. These limitations have greatly hindered effective drug design targeting the adenosine receptors. In order to overcome these limitations, our specific aims include: (1) Develop a new computational method based on recent success of a robust Gaussian accelerated molecular dynamics (GaMD) technique to enable all-atom simulations of protein-protein interactions (PPIs), called ?PPI-GaMD?. (2) Implement PPI-GaMD in widely used open source simulation packages. (3) Test PPI-GaMD on simulations of specific G protein interactions with the A1 and A2A receptors. (4) Apply PPI-GaMD simulations to determine mechanisms of allosteric modulator binding to the A1 and A2A receptors and allosteric modulation of the GPCR?G protein interactions. (5) Validate simulations in vitro by mutagenesis and binding assays and in vivo by cellular functional assays through collaboration with a leading GPCR experimental group. In turn, the simulations will help us to interpret the experimental data at an atomistic level. Our long-term goals are (1) to develop robust computational methodologies to quantitatively characterize biomolecular recognition in disease-associated cellular signaling pathways and (2) to design effective drug molecules targeting important receptors.